In color image processing, color images may be compressed by transforming an RGB representation into a YUV representation, where the Y channel is an intensity expressed as a linear combination of the three color channels and the U and V channels are respective differences between the R and B channels with the Y channel. When the color channels are largely correlated, the Y channel may be sufficient to describe the image and this channel alone may be compressed.